Predictors Of Suicide Attempts Research Articles

Prediction of suicide attempt in a Swedish population-based cohort.

Suicidal behaviors are prevalent public health concerns, and we need to improve our predictive ability to better inform prevention efforts. Using nationwide longitudinal Swedish registers, we included 344,490 males and 323,177 females born 1982-1990 with information on genetic liability and environmental exposures from birth to age 16: perinatal variables, parental psychopathology (suicide attempt, substance use disorder, major depression), family status, socioeconomic difficulties, peers' psychopathology, and school grades. We conducted sex-specific analysis and developed data-driven predictive models including risk factors that occurred between ages 0 and 16 using structural equation modeling. In both females and males, the best-fitting models reveal a complex risk pathway to suicide attempt. In females, the model indicates four direct effects on suicide attempt risk: the occurrence of suicide attempt in parents during childhood (β = 0.159, 95% CI: 0.118; 0.199) and adolescence (β = 0.115, 95% CI: 0.077; 0.153), suicide attempt in peers (β = 0.068, 95% CI: 0.057; 0.079), and low academic achievement (β = 0.166, 95% CI: 0.156; 0.175). In males, aggregate genetic liability for suicide attempt (β = 0.130, 95% CI: 0.111; 0.148), suicide attempt in parents during adolescence (β = 0.099, 95% CI: 0.074; 0.124), suicide attempt in peers (β = 0.118, 95% CI: 0.108; 0.129), and low academic achievement (β = 0.61, 95% CI: 0.152; 0.171) were related to later suicide attempt. These factors also acted as mediators to explain the association between environmental exposures in childhood and later suicide attempt. These findings illustrate sex-specific pathways to suicide attempt by including risk factors that occur during the development. Results highlight the importance of genetic and family environment but also the prominent role of academic achievement.

Suicidal ideation detection on social media using machine learning: A review

Suicide remains a critical global health issue, with over 800,000 deaths annually and profound impacts on individuals and their communities. The increasing role of social media platforms in expressing personal struggles has highlighted the potential of these platforms for early detection of suicidal tendencies. Advances in machine learning and natural language processing (NLP) have revolutionized this detection process, allowing for more accurate identification of suicidal ideation in social media posts. This review examines the integration of machine learning algorithms in suicide detection, including the use of multi-layered neural networks for analyzing textual, audio, and visual data. It explores various machine learning models and their performance in predicting suicide attempts, with a focus on their accuracy, precision, and recall. The review also discusses applications such as real-time monitoring systems and emergency response mechanisms, emphasizing how these technologies can enhance early intervention and support. By surveying current research and identifying challenges and future directions, this review aims to provide a comprehensive understanding of how machine learning can significantly contribute to suicide prevention efforts through effective social media monitoring.

Age-stratified predictions of suicide attempts using machine learning in middle and late adolescence

BackgroundPrevalence of suicidal behaviour increases rapidly in middle to late adolescence. Predicting suicide attempts across different ages would enhance our understanding of how suicidal behaviour manifests in this period of rapid development. This study aimed to develop separate models to predict suicide attempts within a cohort at middle and late adolescence. It also sought to examine differences between the models derived across both developmental stages. MethodsThis study used data from the nationally representative Longitudinal Study of Australian Children (N = 2266). We selected over 700 potential suicide attempt predictors measured via self-report questionnaires, and linked healthcare and education administrative datasets. Logistic regression, random forests, and gradient boosting algorithms were developed to predict suicide attempts across two stages (mid-adolescence: 14–15 years; late adolescence: 18–19 years) using predictors sampled two years prior (mid-adolescence: 12–13 years; late adolescence: 16–17 years). ResultsThe late adolescence models (AUROC = 0.77–0.88, F1-score = 0.22–0.28, Sensitivity = 0.54–0.64) performed better than the mid-adolescence models (AUROC = 0.70–0.76, F1-score = 0.12–0.19, Sensitivity = 0.40–0.64). The most important features for predicting suicide attempts in mid-adolescence were mostly school-related, while the most important features in late adolescence included measures of prior suicidality, psychosocial health, and future plans. ConclusionsTo date, this is the first study to use machine learning models to predict suicide attempts at different ages. Our findings suggest that the optimal suicide risk prediction model differs by stage of adolescence. Future research and interventions should consider that risk presentations can change rapidly during adolescence.

Accuracy and transportability of machine learning models for adolescent suicide prediction with longitudinal clinical records

Machine Learning models trained from real-world data have demonstrated promise in predicting suicide attempts in adolescents. However, their transportability, namely the performance of a model trained on one dataset and applied to different data, is largely unknown, hindering the clinical adoption of these models. Here we developed different machine learning-based suicide prediction models based on real-world data collected in different contexts (inpatient, outpatient, and all encounters) with varying purposes (administrative claims and electronic health records), and compared their cross-data performance. The three datasets used were the All-Payer Claims Database in Connecticut, the Hospital Inpatient Discharge Database in Connecticut, and the Electronic Health Records data provided by the Kansas Health Information Network. We included 285,320 patients among whom we identified 3389 (1.2%) suicide attempters and 66% of the suicide attempters were female. Different machine learning models were evaluated on source datasets where models were trained and then applied to target datasets. More complex models, particularly deep long short-term memory neural network models, did not outperform simpler regularized logistic regression models in terms of both local and transported performance. Transported models exhibited varying performance, showing drops or even improvements compared to their source performance. While they can achieve satisfactory transported performance, they are usually upper-bounded by the best performance of locally developed models, and they can identify additional new cases in target data. Our study uncovers complex transportability patterns and could facilitate the development of suicide prediction models with better performance and generalizability.

Can profiles of behaviors occurring within 48 h of a suicide attempt predict future severity of suicidal thoughts and reattempt?: An examination of hospitalized patients 12 Months post-discharge

Behavioral warning signs (WS) are near-term changes within individuals, which aid in determining imminent risk for suicide attempts. However, those who attempt suicide differ in their engagement of WS, and it is unclear if these differences relate to future risk of suicidal behavior. Using a sample of 132 adults presenting to a hospital following a suicide attempt, the current study sought to determine if differences in engagement in WS for the index attempt prospectively predicted suicide attempt, frequency of ideation, and intensity of suicide ideation 12 months post discharge. Latent class analyses (LCAs) conducted on 6 behaviors (i.e., alcohol use, nightmares, interpersonal negative life events, suicide communication, risky behavior, low sleep, and high sleep) found a 5-class solution optimally fit the data. One identified class, characterized by engagement in risky behaviors the hours before an attempt differed from other identified classes in terms of risk for future suicidal ideation and behaviors. More specifically, participants in “High Risky Behavior” class had higher rates of 12-month suicide reattempt, significantly more frequent suicide ideation, and significantly worse intensity of suicide ideation during the 12 months following their index attempt compared to participants endorsing typical patterns of WS. These results held when adjusting for various traditional baseline covariates (e.g., depressive symptoms). The current study demonstrates that patterns of behavioral WS may be utilized as their own prognostic indicator of future suicidal ideation and behaviors among high-risk individuals reporting a recent suicide attempt, which can inform post-discharge clinical intervention and prevention efforts.

Development and external validation of a logistic and a penalized logistic model using machine-learning techniques to predict suicide attempts: A multicenter prospective cohort study in Korea

Despite previous efforts to build statistical models for predicting the risk of suicidal behavior using machine-learning analysis, a high-accuracy model can lead to overfitting. Furthermore, internal validation cannot completely address this problem. In this study, we created models for predicting the occurrence of suicide attempts among Koreans at high risk of suicide, and we verified these models in an independent cohort. We performed logistic and penalized regression for suicide attempts within 6 months among suicidal ideators and attempters in The Korean Cohort for the Model Predicting a Suicide and Suicide-related Behavior (K-COMPASS). We then validated the models in a test cohort. Our findings indicated that several factors significantly predicted suicide attempts in the models, including young age, suicidal ideation, previous suicidal attempts, anxiety, alcohol abuse, stress, and impulsivity. The area under the curve and positive predictive values were 0.941 and 0.484 after variable selection and 0.751 and 0.084 in the test cohort. The corresponding values for the penalized regression model were 0.943 and 0.524 in the original training cohort and 0.794 and 0.115 in the test cohort. The prediction model constructed through a prospective cohort study of the suicide high-risk group showed satisfactory accuracy even in the test cohort. The accuracy with penalized regression was greater than that with the “classical” logistic model.

The Role of Childhood Trauma, Psychosocial Stress, and Interpersonal Conflicts in the Prediction of Suicide Attempt with the Mediating Factor of Physical Health: A Descriptive Study

The Role of Childhood Trauma, Psychosocial Stress, and Interpersonal Conflicts in the Prediction of Suicide Attempt with the Mediating Factor of Physical Health: A Descriptive Study

Is serum homocysteine level a biomarker of suicide attempts: A preliminary study

Suicide is a global public health concern, and understanding its multifaceted determinants is crucial for effective prevention. This study was designed to find an answer to the question of whether serum homocysteine level can be a biomarker of suicide attempts. This preliminary study involving 90 participants (45 suicide attempt cases and 45 controls) was conducted at Elazig Fethi Sekin City Hospital. Biochemical analyses were performed to assess serum homocysteine, vitamin B12, and folic acid levels. Statistical analyses, including t-tests, Mann-Whitney U tests, and ROC analysis, were employed to explore differences between groups and assess the diagnostic potential of homocysteine. Elevated homocysteine levels were found in individuals who attempted suicide compared to the control group (p= <0.001). Additionally, lower levels of vitamin B12 (p=<0.001) and folic acid (p=<0.001) were observed in the suicide attempt group. ROC analysis indicated a significant diagnostic potential for homocysteine in predicting suicide attempts (AUC = 0.845, sensitivity = 91%, specificity = 71%). This study establishes a significant association between high homocysteine levels and suicide attempts, accompanied by lower vitamin B12 and folic acid levels. The findings suggest a potential link between disturbances in homocysteine metabolism and suicidal tendencies, urging further research to establish causation and explore therapeutic implications. Consideration of the study's limitations and directions for future research are warranted.

Intentions to use mental health and suicide prevention resources among individuals with symptoms of the suicide crisis syndrome and/or suicidal ideation.

The suicide crisis syndrome (SCS) has demonstrated efficacy in predicting suicide attempts, showing potential utility in detecting at-risk individuals who may not be willing to disclose suicidal ideation (SI). The present international study examined differences in intentions to utilize mental health and suicide prevention resources among community-based adults with varying suicide risk (i.e., presence/absence of SCS and/or SI). A sample of 16,934 community-based adults from 13 countries completed measures about the SCS and SI. Mental health and suicide prevention resources were provided to all participants, who indicated their intentions to use these resources. Individuals with SCS (55.7%) were just as likely as those with SI alone (54.0%), and more likely than those with no suicide-related symptoms (45.7%), to report willingness to utilize mental health resources. Those with SI (both with and without SCS) were more likely to seek suicide prevention resources (52.6% and 50.5%, respectively) than those without SI (41.7% and 41.8%); however, when examining endorsements for personal use, those with SCS (21.6%) were more likely to use resources than individuals not at risk (15.1%). These findings provide insight into individuals' willingness to use resources across configurations of explicitly disclosed (SI) and indirect (SCS) suicide risk.

Prediction of Suicide Attempt based on Psychological Pain, Interpersonal Needs, and Peer Pressure in Students

Aim: The aim of this study was to predict suicide attempt based on psychological pains, interpersonal needs, and peer pressure in students. Methods: The research method was descriptive-correlational. The statistical population of the study included all students in city of Rasht (Iran), from which 250 people were selected through random sampling. Research instruments included Schneidmann (1993) Psychache Scale, Joiner et al. (2006) Interpersonal Needs, Dehingra, Budoshk, &amp; O'Connor (2016) Suicide Attempt, and a peer pressure researcher-made questionnaire through which data was collected and analyzed using Pearson correlation and regression. Finding: The results showed that there was a significant relationship between psychological pain (psychache), perceived burdensomeness, and peer pressure with suicide attempts (P &lt;0.01). And the results of regression analysis showed that psych ache and neutral belongingness can predict suicide in students (P &lt;0.05). But the role of psychological pain has a higher power of predictability in suicide attempts among students. Conclusion: It can be concluded that psychological painss and neutral belonging are important factors that can predict suicide in students.

Inhibition and Activation Systems of Behavior and Suicide Mediating Role of Cognitive Emotion Regulation.

This study aimed to predict suicide attempts based on behavioral activation and inhibition systems (BAS/BIS) with the mediating role of cognitive emotion regulation (CER) strategies among adolescents of Abyek City. The research methodology was descriptive-correlational. The cluster sampling method was used, and the resulting sample included 194 adolescents who responded to the BAS/BIS questionnaire (Carver & White, 1994), CER questionnaire, and Beck scale for suicide ideation. Path analysis was employed to evaluate the proposed model using AMOS 24.0 and SPSS 27 softwares. The findings indicate that the proposed model is fitted with the data well. The results of path analysis showed that the BAS/BIS systems explain 27% of the variance of maladaptive CER and 61% of the variance of adaptive CER. In addition, maladaptive CER, BAS, and BIS explain 65% of the variance of suicide attempts. Low levels of BAS, high levels of BIS, and the use of maladaptive CER skills can be considered risk factors for suicide attempts in adolescents.

Predictive Validity of the K-CAT-SS in High-Risk Adolescents and Young Adults

Suicide is a leading cause of death in adolescents and young adults and has increased substantially in the past 15 years. Accurate suicide risk stratification based on rapid screening can help reverse these trends. This study aimed to assess the ability of the Kiddie Computerized Adaptive Test Suicide Scale (K-CAT-SS), a brief computerized adaptive test of suicidality, to predict suicide attempts (SAs) in high-risk youth. A total of 652 participants (age range, 12-24 years), 78% of whom presented with suicidal ideation or behavior, were recruited within 1 month of mental health care contact. The K-CAT-SS, scaled from 0 to 100, was administered at baseline, and participants were assessed at about 1, 3, and 6 months after intake. Weekly incidence of SAs was assessed using the Adolescent Longitudinal Interval Follow-up Evaluation and Columbia-Suicide Severity Rating Scale. A secondary outcome was suicidal behavior, including aborted, interrupted, and actual SAs. The K-CAT-SS showed a 4.91-fold increase in SAs for every 25-point increase in the baseline score (95% CI 2.83-8.52) and a 3.51-fold increase in suicidal behaviors (95% CI 2.32-5.30). These relations persisted following adjustment for prior attempts; demographic variables including age, sex, gender identity, sexual orientation, and race/ethnicity; and other measures of psychopathology. No moderating effects were identified. At 3 months, area under the receiver operating characteristic curve was 0.83 (95% CI 0.72-0.93) for 1 or more SAs. The K-CAT-SS is an excellent tool for suicide risk stratification, particularly in higher-risk populations where other measures have shown lower predictive validity.

Intimate Partner Violence and Suicide Attempt Among Women in the Philippines

Intimate partner violence is among the leading women’s health issues globally. Hence, the primary aim of this study was to determine if intimate partner violence (IPV) against women is a significant risk factor in suicide attempts, and whether such an association is moderated by help-seeking behavior. A predictive correlational design was employed to determine associations between the variables of interest. The study sample consisted of 1,413 currently married or cohabiting Filipino women with exposure to IPV. In the direct effects model, physical IPV, psychological IPV, and sexual IPV emerged as significant risk factors for suicide attempts. In the moderation analysis, help-seeking behavior did not demonstrate a significant interaction with IPV in predicting suicide attempts. The current data highlight the importance of identifying and screening IPV-abused women in communities who may be at risk of attempting suicide. Additionally, suicide prevention strategies must consider the potential role of help-seeking behavior in increasing the risk of suicide among IPV-abused women.

Validation of a Multivariable Model to Predict Suicide Attempt in a Mental Health Intake Sample

Given that suicide rates have been increasing over the past decade and the demand for mental health care is at an all-time high, targeted prevention efforts are needed to identify individuals seeking to initiate mental health outpatient services who are at high risk for suicide. Suicide prediction models have been developed using outpatient mental health encounters, but their performance among intake appointments has not been directly examined. To assess the performance of a predictive model of suicide attempts among individuals seeking to initiate an episode of outpatient mental health care. This prognostic study tested the performance of a previously developed machine learning model designed to predict suicide attempts within 90 days of any mental health outpatient visit. All mental health intake appointments scheduled between January 1, 2012, and April 1, 2022, at Kaiser Permanente Northern California, a large integrated health care delivery system serving over 4.5 million patients, were included. Data were extracted and analyzed from August 9, 2022, to July 31, 2023. Suicide attempts (including completed suicides) within 90 days of the appointment, determined by diagnostic codes and government databases. All predictors were extracted from electronic health records. The study included 1 623 232 scheduled appointments from 835 616 unique patients. There were 2800 scheduled appointments (0.17%) followed by a suicide attempt within 90 days. The mean (SD) age across appointments was 39.7 (15.8) years, and most appointments were for women (1 103 184 [68.0%]). The model had an area under the receiver operating characteristic curve of 0.77 (95% CI, 0.76-0.78), an area under the precision-recall curve of 0.02 (95% CI, 0.02-0.02), an expected calibration error of 0.0012 (95% CI, 0.0011-0.0013), and sensitivities of 37.2% (95% CI, 35.5%-38.9%) and 18.8% (95% CI, 17.3%-20.2%) at specificities of 95% and 99%, respectively. The 10% of appointments at the highest risk level accounted for 48.8% (95% CI, 47.0%-50.6%) of the appointments followed by a suicide attempt. In this prognostic study involving mental health intakes, a previously developed machine learning model of suicide attempts showed good overall classification performance. Implementation research is needed to determine appropriate thresholds and interventions for applying the model in an intake setting to target high-risk cases in a manner that is acceptable to patients and clinicians.

Transdiagnostic inflammatory and oxidative biomarkers with predictive capacity of self-injurious behavior in impulsive and unstable disorders

IntroductionAlterations in inflammatory processes have previously been reported in impulsive and unstable disorders, as well as in other psychiatric conditions. In order to investigate transdiagnostic biomarkers associated with various phenotypic features of these disorders, this study is designed to identify biomarkers of inflammatory and oxidative endophenotypes related to autolytic behavior. MethodsPeripheral blood mononuclear cells were collected from 35 patients with borderline personality disorder (BPD), 29 patients with restrictive eating disorder (rED), 21 patients with purging eating disorder (pED) and 23 control subjects. Plasma levels of different inflammatory and oxidative factors were measured by ELISA and the expression of selected proteins was by Western Blot. Principal component analysis (PCA) was performed to categorize the different inflammatory factors. Additionally, Ancova was performed to observe the differences in the principal components among the different groups and logistic regression analysis was conducted to assess the predictive capacity of these components for autolytic behaviors. ResultsWe found two inflammatory/oxidative components were associated with BPD, characterized by high levels of JNK and ERK and low levels of GPx, SOD and Keap1; and two other inflammatory/oxidative components were linked to pED, associated with more JNK, TBARS and TNF-α and less GPx and SOD. Two components, with more JNK and ERK and less GPx, SOD and Keap1, predicted non-suicidal self-injury and three components, with higher JNK, TBARS and TNF-α levels and lower GPx, SOD and iNOS levels, predicted suicide attempts. ConclusionsThese results strongly support the endophenotypic characterization of impulsivity and the identification of transdiagnostic inflammatory/oxidative biomarkers relevant to autolytic behavior in impulsive and unstable disorders. These dates lay the groundwork for developing of screening tests for these biomarker components to rapidly detect biological risk factors for specific impulse control disorders and future self-injurious behaviors.

Machine Learning-Based Prediction of Suicidality in Adolescents With Allergic Rhinitis: Derivation and Validation in 2 Independent Nationwide Cohorts.

Given the additional risk of suicide-related behaviors in adolescents with allergic rhinitis (AR), it is important to use the growing field of machine learning (ML) to evaluate this risk. This study aims to evaluate the validity and usefulness of an ML model for predicting suicide risk in patients with AR. We used data from 2 independent survey studies, Korea Youth Risk Behavior Web-based Survey (KYRBS; n=299,468) for the original data set and Korea National Health and Nutrition Examination Survey (KNHANES; n=833) for the external validation data set, to predict suicide risks of AR in adolescents aged 13 to 18 years, with 3.45% (10,341/299,468) and 1.4% (12/833) of the patients attempting suicide in the KYRBS and KNHANES studies, respectively. The outcome of interest was the suicide attempt risks. We selected various ML-based models with hyperparameter tuning in the discovery and performed an area under the receiver operating characteristic curve (AUROC) analysis in the train, test, and external validation data. The study data set included 299,468 (KYRBS; original data set) and 833 (KNHANES; external validation data set) patients with AR recruited between 2005 and 2022. The best-performing ML model was the random forest model with a mean AUROC of 84.12% (95% CI 83.98%-84.27%) in the original data set. Applying this result to the external validation data set revealed the best performance among the models, with an AUROC of 89.87% (sensitivity 83.33%, specificity 82.58%, accuracy 82.59%, and balanced accuracy 82.96%). While looking at feature importance, the 5 most important features in predicting suicide attempts in adolescent patients with AR are depression, stress status, academic achievement, age, and alcohol consumption. This study emphasizes the potential of ML models in predicting suicide risks in patients with AR, encouraging further application of these models in other conditions to enhance adolescent health and decrease suicide rates.

Gender differences in risk factors for suicide attempts among young, first-episode and drug-naive major depressive disorder patients with anxiety symptoms.

Suicide attempts and anxiety are common commodities in patients with major depressive disorder (MDD), and suicide attempts are often associated with anxiety symptoms. Studies have found gender differences in several aspects of MDD; however, gender differences in suicide attempts in young first-episode and drug-naive (FEDN) MDD patients with anxiety remain unknown. This study aimed to investigate potential gender differences in the prevalence of suicide attempts and associated risk factors among young FEDN MDD patients with anxiety in a Chinese Han population. A cross-sectional study was conducted on 1289 young patients with FEDN MDD. Demographics, clinical characteristics, and biochemical parameters of patients were collected. Suicide attempters accounted for 23.80% and 26.12% of male and female FEDN MDD patients with anxiety, respectively, with no significant gender differences. Binary logistic regression analyses showed that anxiety, clinical global impression severity, and thyroid peroxidase antibody significantly predicted suicide attempts in both male and female FEDN MDD patients with anxiety, while body mass index significantly predicted suicide attempts only in males, and psychotic symptoms predicted suicide attempts only in females. The present study represents the first large-scale investigation of gender differences in the prevalence of suicide attempts and related risk factors among young FEND MDD patients with anxiety in the Chinese Han population. The results indicate that risk factors associated with suicide attempts vary by gender among young FEND MDD patients with anxiety, although a comparable rate of suicide attempts was observed in both female and male patients.

Predicting suicidal behavior in individuals with depression over 50 years of age: Evidence from the UK biobank.

To construct applicable models suitable for predicting the risk of suicidal behavior among individuals with depression, particularly on the progression from no history of suicidal behavior to suicide attempts, as well as from suicidal ideation to suicide attempts. Based on a prospective cohort from the UK Biobank, a total of 55,139 individuals aged 50 and above with depression were enrolled in the study, among whom 29,528 exhibited suicidal behavior. Specifically, they were divided into control (25,611), suicidal ideation (24,361), and suicide attempt (5167) groups. Least absolute shrinkage and selection operator (LASSO) regression was used to identify a subset of important features for distinguishing suicidal ideation and suicide attempts. We used the Gradient Boosting Decision Tree (GBDT) algorithm with stratified 10-fold cross-validation and grid-search to construct the prediction models for suicidal ideation or suicide attempts. To address the dataset imbalance in classifying suicide attempts, we used random under-sampling. The SHapley Additive exPlanations (SHAP) were used to estimate the important variables in the GBDT model. Significant differences in sociodemographic, economic, lifestyle, and psychological factors were observed across the three groups. Each classifier optimally utilized 8-11 features. Overall, the algorithms predicting suicide attempts demonstrated slightly higher performance than those predicting suicidal ideation. The GBDT classifier achieved the highest accuracy, with AUROC scores of 0.914 for suicide attempts and 0.803 for suicidal ideation. Distinctive predictive factors were identified for each group: while depression's inherent characteristics crucially distinguished the suicidal ideation group from controls, some key predictors, including the age of depression onset and childhood trauma events, were identified for suicide attempts. We established applicable machine learning-based models for predicting suicidal behavior, particularly suicide attempts, in individuals with depression, and clarified the differences in predictors between suicidal ideation and suicide attempts.

Prediction models of suicide and non-fatal suicide attempt after discharge from a psychiatric inpatient stay: A machine learning approach on nationwide Danish registers.

To develop machine learning models capable of predicting suicide and non-fatal suicide attempt as separate outcomes in the first 30 days after discharge from a psychiatric inpatient stay. Prospective cohort study using nationwide Danish registry data. We included individuals who were 18 years or older, and all discharges from psychiatric hospitalizations in Denmark from 1995 to 2018. We trained predictive models using 10-fold cross validation on 80% of the data and did testing on the remaining 20%. The best model for predicting non-fatal suicide attempt was an ensemble of predictions from gradient boosting (XGBoost) and categorical boosting (catBoost). The ROC-AUC for predicting suicide attempt was 0.85 (95% CI: 0.84-0.85). At a risk threshold of 4.36%, positive predictive value (PPV) was 11.0% and sensitivity was 47.2%. The best model for predicting suicide was an ensemble of predictions from random forest, XGBoost and catBoost. For suicide, the ROC-AUC was 0.71 (95% CI: 0.70-0.73). At a risk threshold of 0.15%, PPV was 0.34% and sensitivity was 56.0%. The most contributing predictors differed when predicting suicideand suicide attempt, indicating that separate models are needed. The ensemble model was fair across sex and age, and more so than the penalized logistic regression model. We achieved good performance for predicting suicide attempts and demonstrated a clinical application of ensemble models. Our results indicate a difference in predictive performance for models predicting suicide and suicide attempt, respectively. Thus, we recommend that suicide and suicide attempt are treated as two separate endpoints, in particular for clinical application. We demonstrated that the ensemble model is fairer across sex and age compared with a penalized logistic regression, and therefore we recommend the use of well-tested ensembles despite a more complex explainability.

Comparison of Suicide Attempts and Associated Risk Factors During the COVID-19 Pandemic and One Year Before

Background: Many studies have investigated the rate of suicide attempts and their risk factors during the coronavirus disease 2019 (COVID-19) pandemic; however, the results are still challenging. Objectives: This study investigated and compared the risk factors effective in suicide attempts during the pandemic period and one year before. Methods: This case-control study was conducted at Loghman Hakim hospital, Tehran, Iran. Random sampling included all patients in the inpatient ward who were over 12 years of age and committed suicide during the pandemic within October 1, 2021, to August 30, 2022 (case group; n = 160) or one year before within March 11, 2019, to March 11, 2020 (control group; n = 160). Demographic information, mental-psychological status, and socioeconomic status (SES) of the patients in the case group, before and after the COVID-19 pandemic, were recorded during face-to-face interviews using a predesigned semi-structured questionnaire by study staff trained and supervised by a clinical psychiatrist. Additionally, the data of the patients in the control group were retrospectively recorded during the review of medical records and telephone interviews. The Wilcoxon signed-rank test was used to compare the variables in the case group before and after the COVID-19 pandemic. The Student’s t-test and chi-square test were used to compare the variables between the case and control groups. Additionally, to determine the risk factors predicting suicide attempts during the pandemic, a multivariable logistic regression test was used. The data were analyzed using PASW18 software. Results: The results showed that the patients in the case and control groups had significant differences in terms of age (P = 0.01), underlying mental disorders (P = 0.032), and parental relationship (P = 0.001). Moreover, the variables of suicidal ideation (P = 0.002), feeling lonely (P = 0.023), sleep disorder (P = 0.32), and domestic violence (P = 0.004) in the case group were significantly higher than the control group. In addition, the comparison of these variables in the patients of the case group, before and after the COVID-19 pandemic, showed that the start of the pandemic triggered these factors and led to an increase in suicide attempts in this population. Finally, the logistic regression analysis showed that the variables of SES, having a close friend, parental relationship, previous planning for suicide, domestic violence, and job loss could be considered predictive risk factors for suicide attempts during the pandemic period. Conclusions: According to the present study’s findings, the psychological and socioeconomic conditions triggered by the COVID-19 pandemic led to an increased likelihood of suicidal ideation or attempts in vulnerable individuals.